Computer Science and Programming

A veteran software engineer draws parallels between the Extreme Programming movement of the late 1990s and today's generative AI era, arguing that both represent 'rigor relocation' rather than loss of discipline. Just as XP replaced heavyweight processes with tighter feedback loops, and dynamic languages replaced static types with test-enforced correctness, AI-assisted development demands stricter specification of intent and ruthless evaluation of outputs. The core thesis: probabilistic code generation only works when deterministic constraints exist at the edges. Engineers who thrive will treat generation as a capability requiring more precision in specification, not less, and will build evaluation systems that fail loudly when code drifts from intent.

A practical guide on when to use CSS shorthand properties versus longhand equivalents. The core argument is that readability and intent should drive the decision, not brevity. Covers specific properties like background, padding, margin, animation, transition, grid, border, font, and text-decoration, with concrete examples showing when shorthand helps versus when it obscures meaning. Also introduces CSS logical properties as a more resilient alternative to directional shorthands for internationalization support.

A developer explains their migration from GitHub to a self-hosted Forgejo instance, driven by concerns about digital sovereignty rather than reliability. Key reasons include GitHub's absorption into Microsoft's CoreAI division (losing independent leadership), the April 2026 opt-out flip for Copilot training data, and unresolvable US jurisdictional risk under FISA 702 and the CLOUD Act. The post details the technical architecture: Forgejo v15 LTS on a single NUC with Postgres and Traefik, plus a hardened CI runner using KVM isolation, gVisor, weekly destructive rebuilds, nftables egress filtering, and scope-bound tokens. The Dutch government's choice of Forgejo for code.overheid.nl is cited as institutional validation. Honest trade-offs are covered: loss of GitHub's social graph, Actions ecosystem friction, no Dependabot (replaced by Renovate), and lack of enterprise support.

Microsoft has released the May 2026 servicing updates for .NET and .NET Framework, dated May 12, 2026. The update includes security and non-security fixes, addressing three CVEs (CVE-2026-35433, CVE-2026-32175, CVE-2026-42899). Updated versions include .NET 10.0.8, .NET 9.0.16, and .NET 8.0.27, with corresponding container images and Linux install instructions also refreshed.

A developer spent 7 months vibe-coding a GPU-aware Kubernetes TUI (k10s) with Claude, then archived it after the codebase collapsed under its own weight. The post dissects five concrete failure patterns that emerge from AI-assisted coding without architectural guardrails: AI builds features not architecture (leading to god objects), the god object as default AI artifact, velocity illusion causing scope creep, positional data as a time bomb, and AI mishandling state transitions causing data races. Each tenet includes real code examples from the failed codebase and specific CLAUDE.md/AGENTS.md directives to prevent the same mistakes. The author is rewriting from scratch in Rust, doing architecture design by hand before any AI-generated code.

GitHub's reliability has deteriorated sharply, with 257 incidents tracked between May 2025 and April 2026, including 48 major outages. GitHub Actions alone suffered 57 outages in that period. The root cause, per GitHub's CTO, is the explosive growth of agentic AI workflows demanding 30x the platform's designed capacity. High-profile users like Mitchell Hashimoto (Ghostty) and the Zig project have migrated away. Compounding the scaling crisis are engineering failures like an incomplete feature flag that silently reverted thousands of merged pull requests. Microsoft's absorption of GitHub into its CoreAI org and commercial pressure around Copilot are seen as contributing factors. GitHub has declared an 'availability first' mandate, but community patience is running thin as comparable platforms like GitLab and npm handle the same AI-driven growth without comparable disruptions.

Chrome 148 introduces three notable features: CSS name-only container queries allow querying containers by name without specifying a container-type; lazy loading support is added to video and audio elements via the loading attribute, matching existing behavior for img and iframe; and the Prompt API provides web developers direct access to on-device AI (Gemini Nano) supporting text, image, and audio inputs with response constraints for JSON schema and regex formats.

Supabase has made git-free database branching the default for all projects. Previously available only as a feature preview, dashboard branching lets developers create isolated Postgres branches, make schema changes via the SQL or Table Editor, review a diff, and merge — all without a GitHub integration. Git-based branching remains fully supported and both workflows can coexist. The release also introduces pg-delta, a new schema diffing engine built to replace migra with broader Postgres DDL coverage. Dashboard branching is also the default for branches created via the Supabase MCP server, enabling AI tools to iterate on schemas programmatically without touching git.

The May 2026 Svelte update brings TypeScript 6.0 support in SvelteKit, several improvements to remote functions including breaking changes in 2.56.0, and the experimental release of community add-ons in the Svelte CLI. Notable remote function changes include a new `field.as()` API for default form values,

The May 202transport for richer data types, and a 2026 Svemethod on queries. The CLI now separates026 Svand

packages for a cleaner public API. The community showcase features new apps, UI components, state management libraries, and developer tools built with Svelte.

A Salesforce engineer on the Security Mesh platform increased code coverage by 28% without writing any new tests by restructuring Java data models. The approach involved replacing @Data-annotated mutable classes with immutable Java records and @Value annotations, removing auto-generated boilerplate (getters, setters, utility methods) that inflated coverage denominators without representing real business logic. The Builder pattern was introduced to handle object enrichment while preserving immutability. This reduced total measured lines of code, naturally improving the coverage ratio. The post also discusses how excessive boilerplate harms AI-assisted development tools by consuming context window space and reinforcing false system contracts.

A developer reflects on how they went from mocking Linux power users, vim enthusiasts, and terminal-heavy setups to daily driving Arch Linux, Kitty, and Neovim themselves. The shift came from frustration with sluggish, friction-heavy tools like Windows+WSL and bloated VS Code. The result was a faster, more enjoyable workflow — but also the loss of strong opinions, replaced by the dreaded 'it depends' mindset. The post is a candid, self-aware story about how experience erodes certainty and builds tolerance for trade-offs.